zheorgyan

Lead development for two mission critical services powering postal network: Warehouse and Storehouse, optimizing performance and reducing downtime. Key Contributions: •Introduced Prometheus to track API latency, memory leaks, and microservice health. •Proactively fixed a legacy service memory leak, reducing downtime by 20%. Integrated Zabbix with Grafana for centralized alerts and dashboards, cutting incident detection time by 35% and achieving 99.9% uptime. •PostgreSQL Migration: Redesigned schemas and indexes, slashing query times by 15% and accelerating report generation for operational teams.

Crypto custody / fintech platform Built the Go backend for an enterprise crypto custody platform: wallet provisioning, multi-chain transaction signing and approval, AML screening and bank integrations, with a full audit trail across all operations. Designed a blockchain abstraction layer so each chain's specifics live behind a small set of Go interfaces. Adding a new network means implementing those interfaces, with no changes to core logic. Bitcoin, Ethereum with USDT, and Tron run on this design. Implemented transaction signing where private keys never leave a secured boundary. Keys are held in an HSM over PKCS#11, and the signing service works with any backing store transparently. Added multi-signature so a transaction must be confirmed by several employees before it goes out. Built the approval workflow: a transaction reaches the network only after enough employees with the right roles have signed off, with a full history of who signed and when. Added AML screening. Each address is checked against several external providers for risk, and suspicious addresses are blocked before signing. Providers are interchangeable and cover for each other if one is down. Wrote a background service that polls chains for incoming transactions, runs them through AML, and produces reporting. Services talk over message queues, with part of the data flowing through Kafka. Exposed the API contract as shared protobuf definitions used by both backend and frontend, which keeps the message format in sync across services. Ran the whole distributed system with end-to-end tracing and health checks, brought up with a single Docker Compose command. Stack: Go, gRPC/Protocol Buffers, PostgreSQL, Redis, Kafka, NATS, go-ethereum, btcsuite, gotron-sdk, HSM (PKCS#11), Keycloak/OIDC, OpenTelemetry, TypeScript/React, Docker, GKE, GCP

Development of an orchestration system for RPA robots, optimizing task management and data exchange. Key Responsibilities: •Led the architecture design and implementation of both local and global orchestration components. •Integrated orchestrators with other RPA system components to ensure seamless task management and data exchange. •Optimized system performance, reducing task latency and improving resource utilization across robot networks. •Participated in testing, debugging, and troubleshooting to ensure high system reliability. •Collaborated with analysts and operators to align technical solutions with business needs. Key Contributions: •Implemented a remote management and monitoring feature for bot clusters, allowing operators to manually trigger, pause, or reassign tasks in real-time. This reduced operation execution time by 30% and significantly decreased manual error rates. •Developed a distributed task synchronization mechanism using RabbitMQ and gRPC, improving reliability and task state consistency across local orchestrators. •Designed a real-time dashboard with Blazor Server to visualize robot statuses and task queues, enhancing operator visibility and control. •Introduced performance logging and auto-recovery for faulted tasks, leading to a 25% improvement in system uptime. •Refactored critical parts of the orchestration logic in Golang to handle high-concurrency scenarios, improving throughput under load.